Agent framework / SDK

Runtime for durable state, HTTP access, monitoring, and session storage for self-hosted agent platforms.

Python SDK and AgentOS runtime for self-hosted agent platforms with sessions, memory, and knowledge storage.

Small, composable, self-hosted agent core built on Instructor and Pydantic with validated schemas.

Conversational multi-agent framework providing simple single-agent and multi-agent abstractions for orchestration.

Model-agnostic agent SDK with multi-agent orchestration, MCP support, default tracing, and tool-call hooks.

Azure AI Foundry Agent Service runs managed AI agents with tool orchestration and enterprise controls.

Programmatic access to Claude Code runtime with agent loop, hooks, MCP servers, and granular tool permissions.

Convai enables developers to add conversational AI characters to games and virtual worlds.

Platform for building multi-agent AI automations and collaborative agent workflows.

Databricks Mosaic AI Agent Framework helps teams build, evaluate, deploy, and monitor RAG agents.

Low-code LLM application development platform with visual agent building and workflow capabilities.

Visual agent building platform enabling drag-and-drop creation of LLM applications and agent workflows.

Opinionated agent development kit for building, debugging, and deploying AI agents across five languages.

LLM governance framework with Snowglobe simulator and Guardrails OSS for production monitoring and output validation.

Python library for structured LLM outputs using Pydantic validation for reliable data extraction.

Inworld AI provides tools for creating interactive AI characters and game experiences.

Open-source framework for building LLM applications and AI agents with composable chains and 1,000+ integrations.

Agent harness for long-running workflows with planning, context management, and multi-agent orchestration.

Visual agent building tool providing a drag-and-drop interface for creating LangChain-powered agents.

Open-source conversational AI platform with workflow automation and AI agent node capabilities.

SDK extension connecting agent code to 100+ model providers through a unified OpenAI-compatible interface.

Event-driven agent workflow framework modeling execution as typed event handlers for nested pipelines.

TypeScript-first AI agent framework combining workflow orchestration, memory, and observability with Studio IDE.

Multi-agent orchestration SDK unifying AutoGen and Semantic Kernel with graph-based workflows and type safety.

Moveworks Agent Studio helps agencies build and deploy AI agents that work across enterprise systems.

NVIDIA ACE provides generative AI technologies for interactive digital characters in games.

Lightweight multi-agent workflow SDK for building agent handoffs, tool-calling, and delegation workflows.

OpenAI Computer Use lets agents inspect screen state and perform actions through graphical interfaces.

Python agent framework with structured, validated outputs and dependency injection for type-safe agent development.

Enterprise SDK for embedding AI skills into .NET applications with state management and telemetry.

SGLang is an open-source runtime for fast serving and structured execution of language-model programs.

Minimal code-centric agent framework generating executable Python actions with sandboxed execution.

Swamp helps developers and enterprises build AI agents from reusable behaviors, with API integrations, versioned data storage, and vault-managed credentials.

An AI-native foundation for building, managing, and deploying specialized applications and agents.

TypeScript AI primitives and agent abstractions for building streaming agent interfaces in React and Next.js.

Vertex AI Agent Engine provides managed runtime, sessions, memory, and operations for deployed AI agents.